Unraveling the drivers of community dissimilarity and species extinction in fragmented landscapes

Banks-Leite et al. 2013. Ecology 93(12) 2560-2569. DOI:10.1890/11-2054.1. Unraveling the drivers of community dissimilarity and species extinction in fragmented landscapes

Below, we give our first impressions of this article. Please comment below, or tweet Will or Lynsey (maybe use #pegejc). Think of this as a journal club discussion group!

We both wrote our responses quite rapidly this week, but decided to post them anyway and hopefully start a discussion with our readers. Both of us found this paper thought-provoking and hope to have more coherent thoughts to add later in the week…

Will Pearse

Will Pearse

I picked this paper almost entirely on the basis of figure 1b – the fit looked so good I had to read the paper to find out what it meant. The central claim of this paper is that SAR curves miss changes in species composition when fragmentation takes place, meaning there is a disconnect between species richness and the local extinction of species. The quality of the analysis and the size of the dataset make this a particularly pleasant paper to read, and it reminded me of things (like SLOSS, see below) I haven’t thought about for some time.

The 1970s were a good time to be a conservation biologist: you could still fly to conferences without feeling too bad, and debate was raging about whether to have a Single Large Or Several Small (SLOSS) conservation reserve. This paper is a perfect example of why this debate was never fully resolved: larger fragments may have a higher species richness, but they don’t necessarily contain the same species as smaller fragments. To my mind, this is the clearest demonstration of this effect to date; figure 3d shows more species in larger fragments, but (crucially) there are species present in larger fragments that are absent from smaller fragments, and vice-versa. Going further, a fragment’s surroundings matter too: small fragments in pristine forest resemble larger fragments in near-pristine forest, but are nothing like the smaller fragments in heavily deforested surroundings. Hence all of the figures in this paper are sorted by surrounding forest cover, and then fragment size. Let me say this again: I don’t think I’ve ever seen such neat graphs. Ever.

Which is perhaps something to do with individuals’ range size. I’m no field biologist, but some birds fly quite a long way during the day, and others don’t. This means that different spatial scales of habitat damage are going to be relevant for different birds, and extreme logging of forests might be expected to affect wide-ranging birds first irrespective of the size of individual fragments. Could these kinds of traits, which reflect the use of the surroundings (matrix) by birds, be incorporated into further analyses? Perhaps similar things could be found when comparing among taxa – would insects that disperse only a few meters in their entire lifespan produce as nice graphs as these?

According to the supplementary materials, these fragments are ~40-60 years old, so (in my opinion) these are fairly mature fragments – we’re certainly not seeing the immediate after-effects of fragmentation. Which makes me wonder what those species that specialise in smaller habitat patches represent – what kind of Amazonian species is pre-adapted to small fragments of forest in a sea of deforestation? If they’re not that well-adapted to Amazonian living, and have only come in with deforestation, are they definitely using the forest fragments, and not just passing through? Is it possible to quantify the extent to which particular birds are using a resource? I’ll end on that – if anyone knows how this can be done with mist net data, please let me know. I’ve only done a tiny bit of mist-netting and pit-fall trapping in my time (I was dreadful!) and I often wonder how we’re meant to handle transient passers-by.

Lynsey McInnes

Lynsey McInnes

Typically, I shrike at reading ‘Ecology’ papers – I’m not an ecologist I say to myself. Since joining a department more or less full with population geneticists, it turns out I AM an ecologist, albeit somewhat by accident. Needless to state, Will picked this paper and I went along with it because I like species-area relationships…

That was the first concept to be knocked down, turns out SARs suck at characterising community responses to habitat fragmentation! Ooops.

I found this paper really neat. Very cool, very precise question, amazing data, very sexy, very understandable simulations! I really have to stop myself going to have a play with the R code that does the simulations (also neat that this is provided!).

The rationale for the study and the conclusions they come up with seem robust to me, I am just now wondering whether there are any comparably good datasets where these questions could be asked again? Is such a comprehensive dataset necessary? (Probably yes, right?)

I’ve been a bit lazy and not read the J Appl Ecol. paper where the authors test a bunch of metrics to come up with an adequate one for community composition so I was a bit disappointed that the rationale/robustness/power of the measure chosen wasn’t better explained here. Is it biased in anyway? Does it miss anything?

With my macro-hat on, I wonder if this setup could be used to look at turnover/composition on broader temporal and spatial scales. For example, debate is still raging on ecological limits to diversity, lots of signals point to the diversity-dependence of cladogenesis and this is typically explained by niche filling (diversification slows down as niches get filled), lineages competing for limited resources/niche space. However, how does this really work on broad spatial scales where few lineages within a radiation will actually ever come into physical contact, let alone interact. And perhaps more relevant here, the region in which they diversify is not just one homogeneous blob, but a matrix of different sized habitable units more or less connected to each other. I’m rambling a little, but I think incorporating landscape features in models of spatially-explicit diversification may help in explaining the patterns that we see (I have to think more about this though).

Two more things:

  1. I guess (and Will’s the expert here), on broader scales similar studies have been undertaken under the umbrella of community phylogenetics. I’m still getting a handle on how this study fits with the latter…Also, what would happen if we added a phylogenetic take on compositional turnover? Would it simplify things? Box out traits relevant to persistence in the different fragments?
  2. What does this mean in terms of practical conservation guidelines now that we cannot fall back on the faithful SAR? It sounds to me like we need to get our hands on way more elaborate datasets of species identity and move from there into the traits that the different species bring to the fragments they can persist in. This sounds difficult. Are there any easy work-arounds?

Apologies for a very rambley, not well-thought out response to this article. In short, I really enjoyed it, it made me quite fearful for all those conservation decisions based on species numbers and made me thoughtful on whether these insights and setup could bring clarity to questions typically asked at different scales. Watch this space…



About will.pearse
Ecology / evolutionary biologist

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: